Multi-Arm Bandit Life Enhancement

I’m always trying to do too much. You probably are too.

Whether through insecurity, lack of assurance about my choices, or just wanting more, until recently, I spent way too much time trying to do it all. And failing. Todos would get ticked done but big rocks didn’t move.

During covid and lockdowns, forced to do less, it actually got better. Sure, a lot of productivity books say to do less, but there is a difference between reading it and living it.

So, now I have just 4 things:

  1. Habits
  2. Relationships (friends, family, acquaintances, colleagues)
  3. Enthusiasms (coding/hacking, astrophysics, bioengineering)
  4. Experiments

(notice that my job is not one of them)

These are the things I do regularly. Weekly. Habits I’ve written about in other posts, but systems I use to continue to make sure my foundation is secure and life stuff gets done

Hopefully, the first two are self-explanatory. My enthusiasms are not (quite) professional. I focus on those three. In fact, I now use the Work in Process idea from Kanban so that it’s never more than three. Three might even be too much (someting to reflect on when I take some time off this year.).

The idea is to focus on these solely, and move forward larger projects on them through my habits. Focusing has worked wonders to actually accomplish things (ok, that and perhaps the extra hours per day I gained from working remotely.).

So, what are experiments?


The problem with focus is tunnel vision. You miss important things and new, good stuff always seems to come from the edges or not-what-you’re-focused-on. Both personal and world history tell us so.

Also, novelty and challenge are critical to learning and improving. There’s depth, yes, but we learn and are creative by making connections we hadn’t before. Even for core competencies, most of us plateau after a time and it is usually the introduction of some new connection (at leas tin my experience) that let’s us break through and improve our mastery in some way by trying or doing something we had not done (or couldn’t even see before.

Improvement requires a delicate balance. You need to regularly search for challenges that push you to your edge while continuing to make enough progress to stay motivated. Behaviors need to remain novel in order for them to stay attractive and satisfying. Without variety, we get bored. And boredom is perhaps the greatest villain on the quest for self-improvement. — James Clear, Atomic Habits

So, Experiments are a dedicated, explicit habit to attempt one new thing in your life at set intervals (me: every quarter.). If you focus intently, you need to inject the new into your life to discover new things, avoid monotony, and stay relevant (and avoid boring your friends to death on the same topics). Stuff that may change the trajectory of your life.

It’s my multi-arm bandit approach to life optimization. The idea comes from the standard exploration-exploitation tradeoff dilemma in data science and about how to allocate resources (in this case, your time) to maximize expected gain. Yeah, ok… sounds terrible when you put it that way, I know, but stick with me. The idea is simple.

The setup is simple: once every quarter I try something big and new. I explicitly write these things down and sequence them (for example, in Q4, my goal is to riff off NaNoWriMo and write a bit of a novel every day for a month. Have a SFD done by end of my holidays). Have a written rolling list of four things I want to try every year (and am excited if a little intimidated to try). And these can be anything, just so long as they follow a few simple rules:

1. Completely outside the scope of things I am focusing on

Outside the scope: adding quantum computing to my list (even though I want to understand it) does not count as an Experiment since that is more a matter of me better sequencing my priorities in the coding and hacking focus area. This is a new rule I added in for myself when I noticed I was doubling down in core areas rather than really trying new things.

2. Big enough to matter. Small enough not to screw up.

It should be something which, if completed (see below), gives you a clear signal on success or failure and a lever to add to your life. It can’t be too small. I have lots of “small e” experiments I try as well… toying around with skipping breakfast for ketosis, or adding cinnamon to your coffee. These are not substantial enough to qualify. Changing workouts from evenings to mornings. Attempting something like Morning Pages though (even though it sounds small) as a hack to both get out of my own head and to start a habit of writing every day would be though.

3. Should scare you just a little

The idea of both success (and failure) should scare you a little. This is a good sanity check on if an Experiment is worth your time. It should feel a little uncomfortable. Everyone’s tolerance for risk is different, but the idea here is to push yourself a bit with every experiment. Challenge yourself. Don’t terrify yourself, though. I honestly believe personal growth happens only when we become uncomfortable. It may even be a prerequisite. But getting this balance between challenge and ouright apprehension is important. And be wary of how big a factor social disapproval can be even in niche communities or pursuits. I often find I am more worried about feeling an outsider and being judged harshly (often, very unnecessarily) than the actual pursuit itself.

4. You give yourself explicit permission to fail, judgement and consequence free

And critically linked to the above point about risk: you can fail. You absolutely must give yourself permission to have this go horribly wrong, and it ends up being a story you tell your friends about the crazy time you tried to do Bollywood dancing (don’t ask, but yeah…). In fact, when I am talking to people about things, I often just say “I am trying an experiment… " or, after explaining it to people who look at me incredulously, say “It’s an experiment” as a way of waving off peoples’ reservations, objections, or critical judgements.

It’s an experiment. Experiments fail all the time and the key thing is to learn about yourself from it and riff on the results or modify your next attempt and next experiment.

As a side effect, this willingness to fail consequence-free also helps to make you more resilient to trying new things. Wait and see. It’s an unexpected side effect. I simply was willing to try more and different types of things I may not have considered before.

5. If successful, and incorporated into your life, it should change it in some meaningful way.

Success: suddenly you might have something that could alter the trajectory of your days. And isn’t that what we all want? Some amazing passion that help anchor and enrich our lives? Possibly something amazing in art, to science, a passion project or cause, or a new skill or enthusiasm. If you need to think about it more extremely: how would this change your identity if it was something that you incorporated and eventually mastered? (note: mastery is unneccessary. I’d argue I’m an overly enthusiastic amateur in a couple of areas normally reserved for professionals, and I love doing them that way.)

So, that’s the idea. Focus on your core, but pull the handle hard on that multi-arm bandit once every quarter or so to try something truly new and outside your comfort zone (to 2 or 3 times a year minimum rather than just once a year).

Getting Started

Like all habits, this is mostly about starting and then not getting in your own way.

How do you set yourself up for this well, especially if this idea is appealing, but secretly freaking out all your insecurities because you don’t know what the cool, good looking kids at the dance studio are gonna say when you show up? I feel you.

Here’s a few pointers

  1. Make a List
    Come up with a list of 2-4 things you’ve always wanted to try. Brainstorm a huge list, then cut savagely down to a couple of things you really wanna do (sometimes a why in there is good.).

  2. Sequence it and Schedule It
    Start figuring out when (what quarter or trimester or half) you are going to plan to do the thing, where you’re going to do it, and with whom. Having these basic details in place vastly increases the probability you’ll go through with things.

  3. Have Easy Success Criteria
    A lot of success is just showing up. So, for some things you want to start, do a non-sucky beginner’s thing.

    Wanna run? Grab a couch to 5k app and schedule 2 or 3 runs a week into your cal. Always wanted to do improv? Find an ease-in beginner’s improv course near you. Snowboarding? Plan a ski holiday and schedule a week of on-slope courses with the goal of being able to make it down a green run from the chair by end of week.

  4. Dust yourself off and try again if it all goes wrong
    Sometimes experiments fail but you do not have a result. Running was like this for me. For some reason, I was fine when I got on a business trip, running in the hotel gym for 3 weeks abroad, but progress fell apart when I got home (I blame jetlag) and it needed a reset. So, less a fail or success than the experiment just went wrong. When this happens to you, and it will. Don’t beat yourself up. Just reset the experiment, even if you need go back a few steps. Do not declare failure until you’ve hit some sort of success or result criteria (eg. done it 3 months straight, hit 5k, are up to 2X per week etc etc… ).


OK, so you could boil down this entire post to “try new stuff”, but really it’s more about how to create a system of making sure you’re doing that. Especially as we come out of covid and some things are now possible that weren’t the last 18 months, a good setup is key.

So, focus on your core (and know what that is, the subject of another post) and pull the handle on that random slot machine at a regular interval.

It might change your life.

Let me know what you think about the post on Mastodon @awws or via email . I’d love to hear feedback about your own thoughts, processes, or approaches and what may have worked for you or tweaks to the above. Reasoned opinions on why I might be wrong and what might be even better always welcome.